Yowiehunters Forum

WARNING , this may be extremely boring for some .

I recently collected double woodknocks coming from 2 different directions and not so long ago collected what sounded like a frustrated subject moving around the camera trap at different distances and in different directions , possibly even circling the trap . From the Izotope audio program it's possible to measure the volume of each call , footstep , woodknock and so on . But which direction is it all happening from and how far away is it ?

The recent woodknocks got me thinking a bit more about where this is happening . What if we had a radar of sorts . I began looking for an answer and came up with a microphone array . If we had a microphone array we would be able to show behavior . What if we could also see predatory behavior or avoidance of the trap or even circling . This would add a lot of weight to an otherwise odd piece of audio .
So the idea is to have multiple microphones or audio recorders facing in multiple directions in a 4 sided lunch box . The microphones would be the only part of the recorder sticking out of the box and the box itself would block a lot of noise coming from anywhere else but in front of each recorder . If you stood to the left of the box at 50 metres and called out then the audio recorder facing that direction would catch more of the information than any other recorders .

After deploying the microphone array and retrieving the information it would then be possible to find a call in the audio then measure and compare it to the other 4 audio recorders . If the left and front recorders collected more of the call than the back and right audio recorders then the call must be coming from the front and left of the microphone array . If we were also to measure the difference between the 2 best recordings then we could calculate the exact direction of the call .

What about distance ?

I've spent the last 2 months in the forest trying to get audible measurements that I can use as a baseline to calculate distance from audio recordings . I've recorded myself call out numbers (to simulate vocalisations) , a keyboard with a bass drum (to simulate stomping) and the keyboard playing a snare (to simulate woodknocks) , these noises are repeatable in volume so there won't be any discrepency . The surrounding ambience will also play a part in comparing audio distance from one place to the original baseline . If the ambience in one recording is 20% louder than in the original then the subject should be 20% closer .

The outer ring of my voice chart indicates that at a distance of 100 metres a call should be around -29.44 db . If the call is 20% quieter than this figure then the caller should be 20% further or at 120 metres . If the ambience is 10% quieter than the distance increases by 10% 0r 10 metres to 130 metres .
I've found that the audio recorders themselves are slightly more and less sensitive than each other . I've added 4 seperate audio recordings of ambience together and divided them by four to come up with a repeatable adjustment to each recorder , this will balance them or equalise them .

Another problem is microphone arrays have mono microphones . This is easily overcome by adding both left and right channel together and diving by 2 .
Below are actual measurements of a call I collected . These were the two loudest recordings . The lower the number the louder the volume .

BACK RECORDER LEFT RECORDER
LEFT CHANNEL -39.59 db LEFT CHANNEL -37.92 db
RIGHT CHANNEL -39.43 db RIGHT CHANNEL --35.25 db
----------------------------- --------------------------------
AVERAGE TOTAL -39.51 db AVERAGE TOTAL -36.585 db
ADJUSTMENT -1.45375 db ADJUSTMENT -1.93875 db
= -38.03625 db = -34.64625 db
LOUDEST
The difference between the above calculations is 9.3807%
If the front audio recorder was to be named north , the left recorder named west and so on then the above calculations say that the call came from the south west but 9.3807% closer to the west .
I've had to make 3 100 metre circular charts showing volumes at distances . One chart for stomping , one chart for woodknocks , one chart for vocalisations and one 20 metre chart for bi-pedal walking .
The above figures when calculated show that the call came from the south west of the audio recorder at 116.24 metres . Taking into account that the ambience in this area was 12.9% quieter , then the call came from 129.1 metres .

If one was to collect footage or timelapse pictures in the same direction then it would be advantagous to have this information and would add weight to any existing evidence . If there were activity in a known location then it might also be possible to locate footprints or hair samples .

The video below shows the direction and estimated distance of the call to the center of trap . Calls within 100 metres will be more precise .

Screen shot of voice chart attached . You may notice red numbers on the chart . These numbers indicate an audio anomoly which appears in each chart at different distances . Don't ask me why it is the case but at those distances the audio actually gets louder instead of softer . Because it's different in each chart (snare, bass drum, voice , footsteps) I believe it may come down to harmonics , frequency or something similar . These red numbers may give us a more precise indication of distance if the subject was to move through those distances while vocalising .

Hi Rusty
I know that half wavelengths in frequencies can drown out radio transmissions for example a freq. of 150Khz can be dramatically affected by a freq. of 75Khz if transmitted in a comparable power and area.
Furthermore the acoustics of an area would have a massive affect on measurements aswell ie more rock, water and foliage density. Your baseline measurements would need to be carried out in each location prior to any new study.

But it certainly seems like very sound science and would definitely allow you to hone in on a direction in your camera trap.

Another suggestion to establish any patterns of movement would be to set three or more microphone arrays to further triangulate positions of subjects being recorded. For example your camera trap is the centre on the North,East,South and West axis at same the distance from the trap set up four arrays. This will enhance your ability to pin point the location of the subject and effectively double the distance you are able to record too.

Although potentially a costly set up it would more accurately reflect the position of the subject.

Cheers

Slats wrote:Hi Rusty
I know that half wavelengths in frequencies can drown out radio transmissions for example a freq. of 150Khz can be dramatically affected by a freq. of 75Khz if transmitted in a comparable power and area. Cheers

Yeah right , I have no experience in this field . I see what you mean about foliage density though . I was once in a band . Our sound totally changed when the room was full of people . Empty rooms seem to boom . Sound would travel further in open bushland . If that is the case then the distance calculations would be over estimated in thick bush but the movement and directions shouldn't change too much .

I'd love to have multiple arrays but it would be self defeating . Each footstep , call or wooknock needs to be measured individually . Considering all audio files have to opened at the same time , measured , written down and 5 seperate calculations made of each file which adds up to 20 seperate calculations for one call , then it wouldn't be worth the time . It's the movements which will be more interesting .

At the moment this will be a guide or estimation just like the estimated heights of suspected subjects .

I appreciate your thoughts !

Rusty2 wrote:

Slats wrote:Hi Rusty
I know that half wavelengths in frequencies can drown out radio transmissions for example a freq. of 150Khz can be dramatically affected by a freq. of 75Khz if transmitted in a comparable power and area. Cheers

Yeah right , I have no experience in this field . I see what you mean about foliage density though . I was once in a band . Our sound totally changed when the room was full of people . Empty rooms seem to boom . Sound would travel further in open bushland . If that is the case then the distance calculations would be over estimated in thick bush but the movement and directions shouldn't change too much .

I'd love to have multiple arrays but it would be self defeating . Each footstep , call or wooknock needs to be measured individually . Considering all audio files have to opened at the same time , measured , written down and 5 seperate calculations made of each file which adds up to 20 seperate calculations for one call , then it wouldn't be worth the time . It's the movements which will be more interesting .

At the moment this will be a guide or estimation just like the estimated heights of suspected subjects .

I appreciate your thoughts !

All good mate I've enjoyed your youtube videos!
I learnt a little about soundwaves and frequencies etc when I was in the Army and did my regimental signallers course (unit radio operator).

The density of the foliage will definately affect the volume as can moisture in the ground and air and proximity to rocky outcrops or large bodies of water will affect length the soundwave will travel and also may affect direction with reverberation and echoing.

But I think for the general direction and distance to hone in where to investigate further it will certainly work the science and math seems pretty sound (no pun intended) but it wouldn't hurt to cunduct a baseline test in each new location to firm up the science and make it more acceptable to mainstream.

Rusty, is it possible your sound trap could wirelessly activate another audio recorder, at another location, when a sound is triggered form a certain direction (namely that other location)?

Thank you , I see what you mean .

I suppose if I had lazer measureing binoculars I could have my target tree and do a 50 metre call at front , left , right and behind the trap .

This would give me a percentage difference that I could use to reduce or increase my estimation by . Does that make sense ?

Btw , I do appreciate you sharing this information , it helps me to see when I would otherwise be fumbling around in the dark trying to find my way .


B.R , I suppose it's possible , probably best to have one in each direction anyway but it's so expensive and time consuming as it is . It would certainly be the way to go though .
I need a secretary !!!

Mate I can only imagine the expense and time involved.

But since we're talking 'imagining', indulge me. Imagine if you could use the word ‘radar’ for its real meaning, and the word ‘trap’ for the same.

Using the same detection methods (you described in your post to pinpoint the detection of the sound on a 360dg arc) imagine if every direction could activate another smaller/hidden audio recording device, normally hibernated, set in that direction. The smaller device would have a vastly increased opportunity to gather the sound, ‘real-time’, much, much closer than the larger sound trap that could be 50m away. It'd almost be like virtual battleship.

On the strength of that, imagine if the original audio signal could also activate a video recording device pointed in that specific direction, that would also be small, hidden and hibernated until activated. I know I've mentioned this to you before; tree camera's set high capture a greater field of vision, provide a 3d view of the area and are less noticeable than camera's on the ground (yes, pain to install however), but the potential rewards would be worth it.

As for the sound trap being used more as a ‘trap’, imagine placing it in plain sight, doing very little to camouflage it and scenting it up with ‘human-ness’ (I don’t know, Calvin Klein aftershave or something outlandish like that). Then you’d do your utmost to place the smaller audio and video devices in hidden, camouflaged and scented locations (with these I’d 100% ensure there’s no plastic/metal/human processing smells, then maybe rub eucalyptus oil/grass/or even animal faeces on or around them).

Then, the only object catching their attention, both visually and olfactionally, would be the big audio trap sitting on its own. Definitely increase the chances of them focusing solely on the audio trap, literally used as a decoy, for the satellite audio/visual's to covertly record and capture real-time and at their specific locations.

I know... perfect world. Keep up the good work mate.

Rusty2 wrote:Thank you , I see what you mean .

I suppose if I had lazer measureing binoculars I could have my target tree and do a 50 metre call at front , left , right and behind the trap .

This would give me a percentage difference that I could use to reduce or increase my estimation by . Does that make sense ?

Btw , I do appreciate you sharing this information , it helps me to see when I would otherwise be fumbling around in the dark trying to find my way .


B.R , I suppose it's possible , probably best to have one in each direction anyway but it's so expensive and time consuming as it is . It would certainly be the way to go though .
I need a secretary !!!

Yeah that's basically what I was thinking. I think if you where to orient the camera trap and microphone array N,S E,W for example you could use a compass and a tape measure even, cheaper than rangefinders (unless you already have them), then using a replicable sound in pitch and volume at each 45° point you would be better able to offset the acoustic variables in each location and you would have a control set of sounds for each location aswell. This would keep time in location down and time reviewing the information to as sounds below/in a certain db threshold could be ruled out almost immediately.

Furthermore I think it could be added to further validate your current library of sounds. There's definitely a solid theory/hypothesis you have but it would definately need to be thoroughly documented and processes to remain the same at each location so it could be easily validated.

BR
Rusty's principal is pretty much a reverse radar theory instead of sending noise out he is capturing the noise then calculating the distance and direction much the same as Electronic Warfare in the military. A listening station picks up a frequency and can triangulate with the use of multiple stations or "multiple recorders" distance and direction. You can then target the specific area of subject.
I would think that it could be possible to set up satellite stations but you would possibly miss information due to transmission time wirelessly etc but in a perfect world it would certainly paint a great "3D" image of the reseach location.

Hey B.R. , what your suggesting in a live situation may not exist , but it's a great idea . As for hiding anything I think that may be impossible , they'd sniff it out or watch you putting it there .

If anything , I was thinking more of a 360 degree sound activated video camera (similar to what your thinking) that turns itself towards the loudest sound . That might work but it doesn't exist unless we make it .
My radar is a post radar, after the fact so we can see how they move , how they avoid the trap or predatory behavior .

Hey Slats , yeah , I think at the moment that this method would be equal to being there and hearing it for yourself . You won't know exactly how far away the call is but have a reasonable idea . You won't know exactly where the call came from but have a reasonable idea . All I'd really like to show (if I can) is behavior . If we can show predatory behavior along with the captured audio then that in itself would be enough to add a considerable amount more weight to the recordings .

The technology is already there to see the audio in a live situation but I won't be parting with anymore $$$ .

This link is to some of the latest tech in visualising sound .
https://www.acoustic-camera.com/es.html

Hey Rusty
Yeah it certainly would add some weight. I think in combination with your camera trap you would have a better ability to focus in on that direction/distance.
If it starts to produce good results then you could refine the method and document results (I was googling scientific method for wildlife surveys last night out of curiosity)

I would also wonder if a systematic "grid" appoach in a location would yield more information about behaviour and if different geographical groups used different approach behaviour. I reckon some good information could come from this and start establishing those behaviour patterns which could be applied to other areas of research.

I know what you mean about $$$ I just recently spent $500 on some extra equipment I wanted. Hopefully Santa will bring me some trail cameras.....

Hey Slats , B.R. , here's what I mean about behavior .
The video below shows the position of a barking subject (red light) and a lyre bird (green light)
The main subject barks and the lyrebird instantly responds with a warning whistle .
In the period of 1 minute the barking subject moves uphill behind my trap , possibly to get a peek and the lyrebird moves downhill around about 100 metres away .
The subject barks again in the new position and the lyrebird responds again with 2 warning whistles in it's new position .

If that's not evading a predator I don't know what is . There very well may be 2 lyrebirds but only 1 appears on the audio .
This adds a totaly new dimension to the audio because we can see what's happening .

That's awesome Rusty, it definately shows that the lyre bird has moved to evade something especially with the warning calls. Dingoes don't bark feral dogs might but it wouldn't if it was hunting. It shows that movement you are looking for. Then focus in on those cameras and investigate the area the subject moved through. Hair, prints and audio I like it mate. You might be the one who gets that image/footage backed up with sound.

I know when I'm out bush I try to pay attention to bird calls and warning calls it lets you know where to look and pay attention to for anything that might be a predator.

Is there a particular sound recorder you use or is the software that allows you to calculate the position of the subject and could it be reproduced on cheaper scale? i.e. my sound recorder cost $30 odd and you can get no glow trail cameras for around $110 (I'm an ebay junkie it's a problem)

Hey Slats , I use these audio recorders . https://www.amazon.com/Sony-ICD-AX412-S ... B004M8STL8
They used to be $130 each . If you used these ones you could then use the charts .

Any audio recorder would work for finding direction as long as you had 4 of them facing in 4 different directions and used the same process each time . I know the sony recorders above have the capacity to be attached to phone chargers which allows them to record all week . Not all recorders have this function .

Sony Sound Forge does have a statistics window in the tool box and it will measure highlighted audio but I didn't even know it was there and have never used it . May have to do your homework . Izotope RX seems to have more features as it is a repair tool + you can actually see the call and measure it properly .

Yes , I pay close attention to what the birds are up to as well , had to leave an area the other week because of what the currawongs were doing .

If you have 4 audio recorders you'll have to check to see how far out they are from each other because they won't be recording precisely as each other . Once you establish a known adjustment for each recorder then it's only a matter of using a calculator and measuring the audio on each seperate file .

Thanks Rusty I look forward to seeing the results.

I just need to find a reasonably active area now still doing recon at the moment.

Cheers

Rusty2 wrote:Hey Slats , B.R. , here's what I mean about behavior .
The video below shows the position of a barking subject (red light) and a lyre bird (green light)
The main subject barks and the lyrebird instantly responds with a warning whistle .
In the period of 1 minute the barking subject moves uphill behind my trap , possibly to get a peek and the lyrebird moves downhill around about 100 metres away .
The subject barks again in the new position and the lyrebird responds again with 2 warning whistles in it's new position .

If that's not evading a predator I don't know what is . There very well may be 2 lyrebirds but only 1 appears on the audio .
This adds a totaly new dimension to the audio because we can see what's happening .

Thats very cool Rusty. Another innovation in yowie research. I cant fathom the lingo but seeing the results paints a good picture. Great stuff!

Im guessing if a few researchers get on board with the same gear and programs maybe you will get some interesting reults to compare? Lets hope for your sakes that the Hairys wont catch on!

Hi Rus, There is already a kind of technology that can measure distance with sound, Ive owned a few surround sound systems that auto set up the distance for the viewing area to were the speakers are, in other words they use the speakers in reverse at the set up stage. Eg , one system that had required me to clap once loudly above my head, the system then measured the distance from my clap to each of the surround sound speakers, the other system used a small microphone and listened for the test tone emitted from each speaker then set the required distance from the mic to each speaker. The results were very accurate as you could go into the menu to see the measurements, eg , 5.2m left front speaker, 1.3 m left surround speaker, but if speakers can hear a clap to tell a JVC receiver how far away I am after I clap once, that just gets me thinking that somehow this bit of surround sound technology could be used if it could be obtained. For example, if you had 6-8 small satellite speakers placed around in a 360 degree circle on You’re trap facing out, then if it clapped( just bear with me) the program would tell you witch speaker it was closest to, thus telling you the direction and distance as each other speaker would measure a longer distance, maybe, if the technology could be found, or find out how the surround sound systems do this. Great idea rusty.

Hey Slats , good luck with the recon and thanks for the tech advise .

Hey Yowie bait , it's like acoustic location but based on signal strength rather than triangulation . https://en.wikipedia.org/wiki/Acoustic_location

Hey Hominidhunter , that's interesting , sounds like it's using the same method . I can't for the life of me find some cheap tech that will do the job .

Anytime Rusty
I looked into Izotope and Sound Forge the a pretty hexy bit to much for me atm I just brought a night vision monocular but I did find a free program that apprently does audio repair. It's call Audacity.
I've got some audio to go through from Thursday night so I'll let you know about it once it's done.

Might do a review maybe for others looking to do similar.